Seasonal Induction of Gabaergic Excitation in the Central Mammalian Clock

Seasonal induction of GABAergic excitation in the central mammalian clock

Sahar Farajnia, Tirsa L. E. van Westering, Johanna H. Meijer, and Stephan Michel1

Department of Molecular Cell Biology, Laboratory of Neurophysiology, Leiden University Medical Center, 2300 RC Leiden, The Netherlands

Edited by Joseph S. Takahashi, Howard Hughes Medical Institute, University of Texas Southwestern Medical Center, Dallas, TX, and approved May 19, 2014 (received for review October 24, 2013) The balance between excitation and inhibition is essential for the We suggest that changes in the environment, such as day length, proper function of neuronal networks in the brain. The inhibitory can change the balance between GABAergic excitation and in- neurotransmitter γ-aminobutyric acid (GABA) contributes to the hibition, which may contribute to photoperiod-induced phase network dynamics within the suprachiasmatic nucleus (SCN), which adjustments within the SCN network. is involved in seasonal encoding. We investigated GABAergic activity and observed mainly inhibitory action in SCN neurons of mice Results exposed to a short-day photoperiod. Remarkably, the GABAergic Synaptic Activity Varied Between Different Photoperiods. In the activity in a long-day photoperiod shifts from inhibition toward present study, whole-cell patch clamp recordings were per- excitation. The mechanistic basis for this appears to be a change in formed in SCN neurons of mice exposed to various photoperiods the equilibrium potential of GABA-evoked current. These results em- to estimate the effect of day length on spontaneous postsynap- phasize that environmental conditions can have substantial effects on tic GABAergic currents (sPSC; see Materials and Methods for the function of a key neurotransmitter in the central nervous system. the details of isolating GABAergic inputs from glutamatergic currents). Exposure to a short-day photoperiod [8 h light, 16 h circadian | excitation/inhibition balance | calcium | chloride | NKCC1 darkness (LD8:16)] decreased the frequency of GABAergic sPSCs (3.16 ± 0.52 Hz) compared with exposure to a long-day easonal changes in the photoperiod of the Earth’s temperate photoperiod (LD16:8; 8.56 ± 0.86 Hz; P = 0.0000005; Fig. 1 A NEUROSCIENCE Szones affect the behavior and physiology of many organisms and B) and compared with LD12:12 conditions (7.7 Hz) (14). (1). The central circadian clock, located in the suprachiasmatic In a long-day photoperiod, the sPSC frequency was decreased nucleus (SCN) of the anterior hypothalamus, can adapt to changes during the night (5.61 ± 0.86 Hz) compared with during the day in day length and displays a compressed circadian pattern of (P = 0.018). In contrast, neurons recorded from animals adjusted electrical activity in short winter days and a decompressed pat- to short-day photoperiod showed an increase in frequency at tern in long summer days (2). This pattern is based on a change night (6.92 ± 0.79 Hz) compared with in the daytime (P = 0.0002; in the phase distribution of the activity patterns of individual Fig. 1B). The amplitude of sPSCs did not exhibit significant neurons, which becomes broad in the summer and narrow in the differences between and within any of the groups. winter (3). The mechanisms that mediate and regulate photoperiod-induced phase distributions are currently not known. Excitatory Responses to GABA Require Activation of NKCC1. In The neurotransmitter γ-aminobutyric acid (GABA) is believed whole-cell configuration, the pipette filling solution clamps the − to be involved in the phase adjustment and synchronization of intracellular chloride (Cl ) concentration, which correlates to a − the SCN neuronal network (4, 5). GABA and its receptors are calculated Cl equilibrium potential of (ECl) = −41.25 mV. As expressed in most SCN neurons (6). GABAergic inhibition has a consequence, spontaneous excitatory or inhibitory actions of the been indicated to be important in normal physiological function synaptic events cannot be measured (9). Therefore, we performed within the brain. Alterations in this system (i.e., less inhibition) are shown to cause neurological disorders, such as epilepsy and Significance autism (7, 8). In addition to its classical inhibitory function within the SCN network, GABA has more recently been shown to also The inhibitory neurotransmitter GABA plays an important role act as an excitatory transmitter, although its exact role is un- in many brain circuits involved in anxiety, depression, epilepsy, certain (4, 9–13). To understand the influence of photoperiod on and sleep disorders. GABA is critical for maintaining the bal- GABAergic function, we studied synaptic activity, using patch + ance between excitatory and inhibitory transmission, thereby clamp, and GABAergic responses, using Ca2 imaging techni- ensuring proper neuronal network function. Here, we report ques, in the SCN of mice adjusted to long-day and short-day that entraining mice to a long-day photoperiod can modify the photoperiods. We hypothesized that the narrow, synchronized excitatory/inhibitory balance in the suprachiasmatic neuronal phase distribution of active neurons during short-day photo- network by increasing GABAergic excitation. These data sug- periods would result in increased synaptic activity during the day. gest that day length affects the role of GABA in the circadian Surprisingly, however, exposure to a short-day photoperiod de- clock. This finding is important given the widespread applica- creased the frequency of spontaneous GABAergic synaptic events tion of GABAergic drugs and the general increase in the prev- compared with the long-day photoperiod. alence of sleep disorders. Subsequently, we tested the effect of photoperiod on GABA- 2+ induced excitation in the SCN neuronal network. Ca transients Author contributions: S.F., J.H.M., and S.M. designed research; S.F. and T.L.E.v.W. per- were measured in response to GABA stimulation in long-day formed research; S.F. and T.L.E.v.W. analyzed data; and S.F., T.L.E.v.W., J.H.M., and S.M. and short-day photoperiods. Interestingly, of all cells from the wrote the paper. long-day photoperiod, 40% were excitatory and 36% were in- The authors declare no conflict of interest. hibitory. In contrast, in the short-day photoperiod, 28% of the This article is a PNAS Direct Submission. cells were excitatory and 52% were inhibitory. Using perforated Freely available online through the PNAS open access option. patch recordings, we demonstrate that the underlying mecha- 1To whom correspondence should be addressed. E-mail: [email protected]. nism for long-day-induced GABAergic excitation is a depolariz- This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. ing shift in chloride equilibrium potential. 1073/pnas.1319820111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1319820111 PNAS Early Edition | 1of6 Downloaded by guest on October 2, 2021 spatial differences in GABAergic responses we report may be a result of anatomical differences between species, with a more defined SCN subregion in the rat compared with the mouse. + The GABA-mediated Ca2 responses revealed a striking difference in the ratio of excitatory to inhibitory GABAergic activity between animals entrained to long-day versus short-day photoperiods (Fig. 3). Compared with a long-day photoperiod, adaptation to a short-day photoperiod led to significantly more inhibitory responses (36% vs. 52%, respectively; P = 0.0002) and fewer excitatory responses (40% vs. 28%, respectively; P = 0.0008). This difference was even greater in a subset of data recorded within 1 h of midday (Fig. 3D). Under LD12:12 conditions, the percentages of GABA-mediated excitation (32%) and inhibition (43%) were intermediate between their respective values for long-day and short-day photoperiods (Fig. 3E and Fig. Fig. 1. Long-day photoperiod increases frequency of GABAergic sPSCs. (A) S1B). Furthermore, this photoperiodic effect was restricted to Examples of sPSC daytime recordings from SCN neurons of mice entrained to a long-day or short-day photoperiod. The neurons were voltage-clamped the day, as GABA-evoked responses during the night did not at −70 mV. (Scale bars, 50 pA, 1 s.) (B) Mean ± SEM of the frequency of differ between long-day and short-day photoperiod, either in GABAergic sPSCs recorded during the day and night for long-day (8.56 ± excitation (50% vs. 40%, respectively; P = 0.071) or in inhibition 0.86 Hz, n = 63; 5.61 ± 0.86 Hz, n = 50) and short-day (3.16 ± 0.52 Hz, n = 52; (20% vs. 25%, respectively; P = 0.121) (Fig. S2). + 6.92 ± 0.79 Hz, n = 48) photoperiods. GABAergic events were selected from Although GABA-induced Ca2 transients can depend on the raw dataset exemplified in A based on their characteristic decay time 2+ 2+ baseline [Ca ]i (11), we found no difference in baseline [Ca ]i (see Materials and Methods). *P < 0.05, **P < 0.01, ***P < 0.001, in- of SCN neurons during the day between long-day and short-day ’ dependent Student s t test. photoperiods (Fig. S3A). Moreover, we found no correlation between the response type (i.e., excitatory or inhibitory) and 2+ baseline [Ca ]i in either long-day or short-day photoperiods gramicidin perforated patch recordings to measure post- + B Left 2 synaptic potentials in SCN neurons of long-day-entrained (Fig. S3 , ). During the night, baseline [Ca ]i was higher in inhibitory versus excitatory cells only in a long-day photoperiod mice without manipulating E (Fig. 2A).Themajorityof + Cl (P = 0.001). However, no difference was found in baseline [Ca2 ] neurons recorded (92%, n = 12 cells) generated excitatory i of inhibitory and excitatory cells between long-day and short-day postsynaptic potentials (EPSPs), and the frequency of these photoperiods (Fig. S3B, Right). In conclusion, our data on photo- EPSPs was reduced by 71% in the presence of the GABA A periodic modulation of the GABAergic responses were not con- receptor blocker gabazine (P = 0.004; Fig. 2B). Similar re- + founded by a change in baseline [Ca2 ] . cordings in animals entrained to LD12:12 revealed that only i = 21% of the neurons had EPSPs (n 14 cells). Likewise, Long Photoperiod Shifts Equilibrium Potential of GABA-Evoked Current. neuronal activity was affected by gabazine. On blocking GABAA- To further investigate the mechanisms underlying these photope- mediated transmission in long-day-entrained animals, action po- riodic changes in GABAergic signaling, we recorded GABA-evoked tential frequency was increased in 2 of 12 neurons; five neurons had currents using the gramicidin perforated patch technique in the no change in action potential frequency, and the remaining five cells presence of blockers for glutamatergic receptors, calcium channels, = had a significant decrease in frequency (60%; P 0.043). These and action potentials (Fig. 4A). The majority of long-day adapted findings suggest that GABAergic transmission can be excitatory in cells displayed GABA-induced inward currents (eight of nine cells, the SCN and seems to be increased after entrainment to long days. 2+ four animals), whereas more than half of the short-day adapted Previous studies showed that Ca transients are a good ap- cells responded with outward current to GABA application (9 of proximation of neuronal activity (11, 15), with elevations and 16 cells, four animals). Current-clamp recordings in the same cells 2+ 2+ reductions in intracellular Ca concentration ([Ca ]i) reflecting showed that inward currents were always correlated to spontaneous excitation and inhibition, respectively. GABA-mediated excita- − EPSPs, confirming the excitatory nature of endogenous GABAergic tion is based on a change in the activity of the Cl cotransporter activity (Fig. 4B). Because excitatory GABAergic responses seem NKCC1, which can be blocked by bumetanide (9). In the majority to depend on activation of NKCC1 (Fig. 2C) (9, 11), we predicted − of slices from mice entrained to either long-day or short-day a photoperiod-induced change in intracellular Cl concentration photoperiods, we recorded a combination of transient increases and a shift in the equilibrium potential of GABA-evoked current 2+ and decreases in [Ca ]i in response to GABA (Fig. S1A). To (EGABA). Using voltage ramp protocols in the same cells, we re- 2+ confirm that an increase in [Ca ]i is based on the excitatory corded GABA-mediated currents and found that EGABA was sig- action of GABA, we applied bumetanide (10 μM, 5–7min)toSCN nificantly depolarized in long-day (−40.3 mV) compared with neurons from mice adjusted to long-day photoperiod. After short-day (−48.5 mV; P = 0.002) adapted cells (Fig. 4 C and bumetanide application, the amplitude of GABA-evoked eleva- D). On the basis of these values, the estimated intracellular 2+ − tion in [Ca ]i was diminished (P = 0.0003; Fig. 2 C and D). We Cl concentration in the SCN neurons recorded was higher in + therefore conclude that the Ca2 transients we recorded in response the long-day compared with the short-day photoperiod (25 and to GABA application indeed represent neuronal excitation. 18 mM, respectively).

GABAergic Excitations Were Increased in a Long-Day Photoperiod. As Discussion GABA-mediated excitation was previously shown to vary be- Here, we provide compelling evidence that exposure to a long- tween different SCN areas of the rat (9, 11), we first tested the day photoperiod switches the polarity of GABAergic activity in GABA-evoked excitatory and inhibitory responses for regional most SCN neurons from inhibitory to excitatory. Presynaptically, differences. The responses we measured in the mouse SCN were sPSC frequency changes with differing day lengths, whereas post- not significantly different between anterior, middle, and poste- synaptically, the photoperiod affects GABAergic activity within the rior hypothalamic slices (one-way ANOVA with post hoc Bon- SCN by changing the equilibrium potential of GABA-evoked cur- ferroni test) or between dorsal and ventral SCN (χ2-test, two- rent. The increase in excitatory GABAergic activity was reduced − tailed), so we pooled the data for further analysis. The lack of after blocking the Cl cotransporter NKCC1 using bumetanide,

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1319820111 Farajnia et al. Downloaded by guest on October 2, 2021 Fig. 2. GABA-mediated neuronal excitation in the SCN. (A) Example recordings of three SCN neurons adjusted to a long-day photoperiod, depicting EPSPs

(marked by asterisks) and action potentials before (Left) and after (Right) the application of the GABAA receptor blocker gabazine (10 μM, 5 min); note that gabazine eliminated the majority of the EPSPs. Dashed lines indicate the resting membrane potential before and after the treatment. (Scale bars, 20 mV, 1 s.) + (B) Mean ± SEM of EPSP frequency before and after gabazine application (n = 12). **P < 0.01, Wilcoxon signed-rank test. (C) Example Ca2 -imaging NEUROSCIENCE recordings of SCN neurons adjusted to a long-day photoperiod before (left traces) and after (right traces) application of the NKCC1 antagonist bumetanide + (10 μM, 5 min). Bumetanide attenuated the GABA-induced transient Ca2 elevations. (Scale bars, 20 nM, 20 s.) (D) Mean ± SEM of the peak of GABA-induced + Ca2 transients before and after bumetanide application (n = 39). ***P < 0.001, Wilcoxon signed-rank test.

suggesting a modulation of NKCC1 activity or expression. Thus, our suggests a differential role of GABA during day and night, with data show that environmental conditions affect GABAergic activity an increased excitatory action during long days. However, the by modulating cellular properties on a basic biophysical level. requirement of GABA during photoperiodic entrainment and The key mechanisms that contribute to the degree of syn- the role of photoperiod-induced GABAergic excitation in de- chronization within the SCN, reflected in the photoperiodic-in- termining the phase of SCN neurons needs to be addressed in duced changes in phase distribution (3), may depend on the ratio future studies. of excitatory to inhibitory GABAergic activity within the SCN, Changes in the photoperiod were recently reported to affect rather than an overall increase in GABAergic tone. The role of the relative expression of two neurotransmitters in hypothalamic inhibition in synchronization has been shown previously in other neurons after exposure to long days (20). Here, we show that the neuronal networks (16, 17). Whether inhibition would also in- action of a single transmitter system is affected by environmental duce phase synchrony in the SCN remains to be established. In conditions. On the basis of our findings, it would be interesting to the present study, we found a relatively high percentage of in- investigate the influence of photoperiod on neurotransmitter hibition in the short-day compared with the long-day photope- systems in other brain regions. It is worth noting that our study riod, which could contribute to the phase synchrony seen in short has been performed in C57BL/6 mice. Although these mice are days. Freeman et al., however, suggested that GABAergic ac- not seasonal breeders, they clearly show adaptation of behavior tivity could be a phase desynchronizer and destabilizer in the and SCN activity to changes in photoperiod (3). They exhibit SCN (18), but they did not distinguish between GABAergic after-effects in constant darkness, such as changes in rest and excitation and inhibition. We suggest that elevated GABA- activity duration, and in waveform of the SCN rhythm, indicating mediated excitation during the long-day photoperiod could de- a “memory” for day length. stabilize the phase of neuronal activity and allow phase dispersal. The balance between excitation and inhibition has been The relation between GABAergic excitation and inhibition proven to be necessary to preserve a normal physiological may thus determine the photoperiod-induced phase distribu- function in the central nervous system (7, 8). GABA, specifically, tion in the SCN network. As such, it is plausible that both plays a pivotal role in the maintenance of this balance, as it is the actions of GABA [phase-setting (4) and destabilization (18)] major inhibitory neurotransmitter in the brain. During early can facilitate the adjustment of different phase distributions development, GABA is an excitatory transmitter. A disturbance during various photoperiods. in GABAergic function at this stage causes neurodevelopmental The contribution of GABAergic signaling to photoperiodic disorders, such as mental retardation, Angelman syndrome, ep- entrainment seems to depend on the state of synchrony within ilepsy, and autism (21). Recently, bumetanide has been shown the SCN network and may be restricted to the day. After pho- to improve behavior in children with autism by reducing toperiodic entrainment to a regime of LD20:4, the SCN neurons GABAergic excitation (22). GABA also acts as an excitatory require neurotransmission by GABA as well as vasoactive neurotransmitter in the mature but pathological brain (23). intestinal peptide to readjust the phase of their rhythms in Remarkably, in the mature healthy nervous system, GABA can PERIOD2 expression (19). We have shown in the present study also act as an excitatory transmitter in cortex, hippocampus, and that the effect of a long-day photoperiod (LD16:8) on the ratio amygdala in addition to the SCN (24). Although the role of between excitation and inhibition is restricted to the day. This GABAergic excitation is not known in the mature brain, it is

Farajnia et al. PNAS Early Edition | 3of6 Downloaded by guest on October 2, 2021 Fig. 3. Increased GABA-mediated excitation in the long-day photoperiod. (A) Examples of fura-2-AM loaded SCN neurons from mice entrained to a long-day (Left) or short-day (Right) photoperiod. The numbers indicate the neurons that are shown in B. Color scale indicates fluorescence intensity at 380 nm exci- + tation in arbitrary units. (B) GABA-induced excitatory (orange traces) and inhibitory (blue traces) Ca2 transients recorded during the day from the cells indicated in A. (Scale bars, 50 nM, 10 s.) (C) Pie charts depicting the distributions of response types of the cells on GABA stimulation for long-day (n = 760) and short-day (n = 680) photoperiod-entrained neurons. (D) Summary of the percentages of excitatory and inhibitory responses in long-day (n = 222) versus short- day (n = 218) photoperiod-entrained neurons recorded within 1 h around midday. *P < 0.05, **P < 0.01, ***P < 0.0001, χ2-test. (E) Summary of the ratios of excitatory to inhibitory GABAergic signaling in different photoperiods.

clear that the balance of excitation and inhibition is regulated given light schedule. Experiments were performed within a 3-h interval within a narrow range, and deviations from this range can lead to centered in the middle of the day and the middle of the night. We used neuropsychiatric disorders (25). external time (ExT) to allow for easier comparison between different pho- One form of depression that is susceptible to changes in day toperiods. ExT 12 is defined as midday (middle of the light period), and ExT 0 as midnight (middle of the dark period) (30). All experimental procedures length is seasonal affective disorder, which is at least partially were approved by the Committee on Animal Health and Care of the Dutch based on photoperiodic alterations of the circadian system (26). government (no. 11010). A recent study using a day-active rodent model suggests a link between depressive-like behaviors and the photoperiodic re- Slice Preparation. The mice were killed between ExT 8–9 for the daytime sponsiveness of the circadian clock (27). Although there is still recordings. For nighttime recordings, animals adjusted to short-day photo- insufficient evidence on the neurobiological mechanisms un- period were killed between ExT 11–12, whereas animals adjusted to long- – derlying seasonal affective disorder, altered GABAergic signal- day photoperiod were killed between ExT 15 16. Animals entrained to LD12:12 were killed between ExT 6–7. ing in the SCN neuronal network should be considered a potential Hypothalamic slices containing the SCN were prepared as described pre- contributing factor. Furthermore, considering that GABA is a viously (31). In brief, brains were quickly removed and submerged into target for therapeutic treatments of sleep disorders and the role modified ice-cold artificial cerebrospinal fluid (ACSF) containing (in mM):

of the SCN in sleep regulating processes, the seasonal influence on NaCl 116.4, KCl 5.4, NaH2PO4 1.0, MgSO4 0.8, CaCl2 1.0, MgCl2 4.0, NaHCO3 GABAergic function may affect the therapeutic manipulation of 23.8, glucose 15, and 5 mg/L gentamicin (Sigma Aldrich, Munich, Germany) – – this neurotransmitter system differently in summer compared with saturated with 95% O2-5% CO2 (pH, 7.2 7.4; osmolality, 290 310 mOsm). winter. Together, these findings imply that some crucial neuronal Coronal slices were cut using a vibratome (VT 1000S, Leica Microsystems, networks within the CNS are sensitive to changes in day length, or Wetzlar, Germany) and subsequently maintained in regular ASCF (CaCl2 increased to 2 mM and MgCl2 decreased to 0 mM) for at least 1 h before even to prolonged artificial light exposure common in modern recordings. Slices containing the SCN were transferred to a recording society (28, 29). chamber (RC-26G, Warner Instruments, Hamden, CT) mounted on the fixed stage of an upright fluorescence microscope (Axioskop 2-FS Plus; Carl Zeiss Materials and Methods Microimaging, Oberkochen, Germany) and constantly perfused with oxy- Animals and Housing Conditions. Male C57BL/6 mice (Harlan, Horst, the genated ACSF (2.5 mL/min). Netherlands; 8–16 wk old; n = 88) were housed under different light regimes, such as long-day (LD16:8), short-day (LD8:16), or equinoctial (LD12:12) Patch Clamp Recordings. Whole-cell voltage-clamp recordings of sPSCs (n = 64 photoperiods, in climate-controlled cabinets with ad libitum access to food slices from 35 animals) were performed using a patch amplifier (EPC 10–2; and water. Before experimentation, the mice were exposed to their re- HEKA, Lambrecht/Pfalz, Germany), as described previously (31). Micropipettes spective photoperiod for a minimum of 4 wk to ensure entrainment to the (with a tip resistance of 5–7MΩ when filled with internal pipette solution)

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1319820111 Farajnia et al. Downloaded by guest on October 2, 2021 NEUROSCIENCE

Fig. 4. EGABA is more depolarized in long-day photoperiod. (A) Example traces of GABA-evoked currents at holding voltage of −50 mV recorded as inward currents in cells adjusted to long-day and as outward currents in cells adjusted to short-day photoperiod. (Scale bar, 5 pA, 5 s.) (B) Examples of excitatory and inhibitory postsynaptic potentials recorded from long-day and short-day photoperiod, respectively, in current clamp configuration. Numbers indicate the resting membrane potential. (Scale bars, 10 mV, 500 ms.) (C) Examples of current responses to a voltage ramp (green line) recorded from cells adjusted to long- and short-day photoperiods in the absence and presence of GABA (Scale bars, 20 mV, 50 ms). The value of ramp potential at which the control and

GABA current traces cross (arrow) is equivalent to EGABA. (Scale bar, Left,20pA50ms;Right, 80 pA, 50 ms.) Inset shows expanded traces of boxed regions within the crossing area. (Scale bars, 10 pA, 10 ms.) (D) Mean EGABA ± SEM recorded from long-day (n = 12) and short-day photoperiod (n = 16). **P < 0.01, independent Student’s t test.

were fabricated from borosilicate glass capillaries using a PC-10 puller (Narishige, potential (16 mV) and corrected the values of membrane potential for the London, United Kingdom) and filled with an internal solution (pH, 7.2–7.3; os- current-clamp recordings. molality, 290–300 mOsm) containing (in mM): 112.5 K-gluconate, 1 EGTA, 10 To determine the time course and equilibrium potential of GABA-evoked Hepes (Na+ salt), 5 MgATP, 1 GTP, 0.1 leupeptin, 10 phosphocreatine, 4 NaCl, currents (EGABA), gramicidin perforated patch voltage-clamp recordings were 17.5 KCl, 0.5 CaCl2, and 1 MgCl2. performed (eight slices from eight animals) and GABA (200 μM, 6 s) was ap- Spontaneous postsynaptic currents were analyzed using the MiniAnalysis plied locally, using a focal application system (ALA-VM8; ALA Scientific program (Synaptosoft, Decatur, GA). Events were sorted by their decay time Instruments, Farmingdale, NY). AMPA and NMDA receptors and Na+ and Ca2+ into glutamatergic PSCs [average decay time, 4.6 ms (32)] and GABAergic channels were blocked using CNQX (25 μM), AP-5 (50 μM), TTX (0.5 μM), and PSCs [average decay time, 15.4 ms (14)]. The GABA receptor antagonist cadmium (25 μM), respectively. Voltage ramps from −100 to 0 mV in 200 ms gabazine (10 μM) was applied after the recording to confirm that the cur- were applied from a holding potential of −50 mV in the absence and sub- rents were GABAergic. Recordings with series resistance ≥40 MΩ were ex- sequent presence of GABA with 40-s intervals (c.f. refs 33 and 34). In 4 ± 0.31 s cluded from the data pool and subsequent analyses. (n = 26), the GABA current (either inward or outward) could reach its maxi- To measure GABA-evoked sPSPs, gramicidin perforated patch current- mum. Therefore, we applied GABA for 6 s to ensure a proper response during clamp recordings were performed (c.f. ref. 9; n = 16 slices from six animals). − the ramp protocol. Control ramp was repeated 40 s after the GABA ramp to GABAergic responses are mediated by transmembrane flux of Cl , and the − check the reversibility of the GABA-evoked current. The membrane potential polarity depends on the Cl equilibrium potential (ECl) of the cell. Unlike at which the current traces recorded in the absence and presence of GABA whole-cell recordings, the gramicidin perforated patch technique does not − change the intracellular Cl concentration, and therefore GABAergic EPSPs crossed was considered as EGABA.ThisvalueforEGABA was corrected for series resistance voltage error by subtracting the product of the net membrane can be measured at unaltered ECl. First, the tip of the pipette was filled with gramicidin-free solution containing (in mM): 143 K-gluconate, 2 KCl, 0.5 current flowing at the cross point and the residual series resistance. EGTA, and 10 Hepes, (pH, 7.2; osmolality, 295 mOsm) to facilitate the for- 2+ 2+ mation of a giga-Ohm seal. The pipettes were then back-filled with the same Ca Imaging. Ca measurements were performed as described previously 2+ solution containing 50 μg/mL gramicidin. Within 6–20 min of forming a giga- (35). Brain slices that included the SCN were loaded with the ratiometric Ca Ohm seal, stable access resistance (40–200 MΩ) was achieved and remained indicator dye fura-2-acetoxymethyl ester (Fura-2-AM). The slices were throughout the recordings. Gabazine (10 μM) was applied to confirm that loaded with 7 μM Fura-2-AM in ACSF at 37 °C for 10 min. The slices were recorded EPSPs are evoked by GABA. We measured the liquid junction then rinsed with freshly oxygenated ACSF for 10–30 min before recording. A

Farajnia et al. PNAS Early Edition | 5of6 Downloaded by guest on October 2, 2021 monochromator (Polychrome V; TILL Photonics, Gräfeling, Germany) was were analyzed using TILLvisION, and the responses of neurons were ana- + used to deliver paired 50-ms light pulses of two excitation wavelengths (340 lyzed using IGOR Pro. Ca2 transients with an increase in amplitude ≥10% and 380 nM). Emitted light (505 nM) was detected by a cooled CCD camera from the baseline level were defined as excitatory, and transients with (Sensicam; TILL Photonics), and images were acquired at 2-s intervals a decrease in amplitude ≥10% were defined as inhibitory. Cells that respon- (0.5 Hz). Single-wavelength images were background subtracted, and ratio ded with both excitatory and inhibitory responses after a single GABA pulse images (340/380) were generated. Region-of-interest-defined cells and mean were defined as biphasic. Last, cells that responded with a change in ampli- + ratio values were determined, from which the intracellular Ca2 concen- tude <10% from baseline were defined as nonresponding. The effects of tration was calculated. Experiments were controlled by imaging software bumetanide were analyzed by inspecting the difference in amplitude of the (TILLvision; TILL Photonics). cell’s response to a GABA pulse before and after bumetanide application. The 2+ GABA (200 μM, 10 s) was applied locally to trigger Ca transients in SCN amplitude of the response was calculated by subtracting the baseline from + neurons, using a focal application system (ALA-VM8). In addition, ACSF the peak Ca2 level achieved during GABA application. + “ ” containing elevated K ( High-K, 20 mM, 10 s) was applied to identify Statistical analyses were performed using SPSS (IBM, Armonk, NY). The = healthy, responding cells (LD16:8: day, n 17 slices from eight animals; appropriate statistical test was selected after using Shapiro-Wilk and Levene’s = = night, n 11 slices from six animals; LD8:16: day, n 16 slices from eight tests to evaluate the normality of the data and homogeneity of variances = = animals; night, n 13 slices from seven animals; LD12:12: day, n 8 slices respectively. All values obtained from the whole-cell patch clamp recordings from five animals). Additional experiments were performed in slices from were tested for significance using an independent Student’s t test (two- μ mice adjusted to a long-day photoperiod to which bumetanide (10 M), tailed). The gramicidin perforated patch recordings were analyzed using the a pharmacological agent that blocks the activity of the NKCC1 cotransporter, nonparametric Wilcoxon signed-rank test (two-tailed). The distributions of – = was administered for 5 7 min before GABA application (n 8 slices from GABA-induced responses were analyzed using the χ2-test (two-tailed). The five animals). effect of bumetanide was measured using the Wilcoxon signed-rank test (two-tailed). The difference in equilibrium potential of GABA-evoked cur- Chemicals. GABA, gramicidin, and all salts were purchased from Sigma- rent in various photoperiods was analyzed using independent Student’s t Aldrich. Gabazine, bumetanide, CNQX, AP-5, and TTX were purchased from test (two-tailed). Differences with P ≤ 0.05 were considered to be significant. Tocris Bioscience (Bristol, United Kingdom), and Fura-2-AM was purchased from Teflabs (Austin, TX). ACKNOWLEDGMENTS. We thank Heleen Post-van Engeldorp Gastelaars and Kit-Yi Yam for their technical assistance. This work was supported by The Data Analysis. Data were collected and analyzed using FitMaster (HEKA), Igor Netherlands Organisation for Scientific Research/Netherlands Organisation Pro (Wavemetrics, Portland, OR), and MiniAnalysis (Synaptosoft). Ca2+ images for Health Research and Development Grant TOPGo 91210064 (to J.H.M.).

1. Walton JC, Weil ZM, Nelson RJ (2011) Influence of photoperiod on hormones, be- 19. Evans JA, Leise TL, Castanon-Cervantes O, Davidson AJ (2013) Dynamic interactions havior, and immune function. Front Neuroendocrinol 32(3):303–319. mediated by nonredundant signaling mechanisms couple circadian clock neurons. 2. Mrugala M, Zlomanczuk P, Jagota A, Schwartz WJ (2000) Rhythmic multiunit neural Neuron 80(4):973–983. activity in slices of hamster suprachiasmatic nucleus reflect prior photoperiod. Am J 20. Dulcis D, Jamshidi P, Leutgeb S, Spitzer NC (2013) Neurotransmitter switching in the Physiol Regul Integr Comp Physiol 278(4):R987–R994. adult brain regulates behavior. Science 340(6131):449–453. 3. VanderLeest HT, et al. (2007) Seasonal encoding by the circadian pacemaker of the 21. Pizzarelli R, Cherubini E (2011) Alterations of GABAergic signaling in autism spectrum SCN. Curr Biol 17(5):468–473. disorders. Neural Plast 2011:297153. 4. Albus H, Vansteensel MJ, Michel S, Block GD, Meijer JH (2005) A GABAergic mecha- 22. Lemonnier E, et al. (2012) A randomised controlled trial of bumetanide in the nism is necessary for coupling dissociable ventral and dorsal regional oscillators within treatment of autism in children. Transl Psychiatr 2:e202. the circadian clock. Curr Biol 15(10):886–893. 23. Ben-Ari Y, Holmes GL (2005) The multiple facets of gamma-aminobutyric acid dys- 5. Liu C, Reppert SM (2000) GABA synchronizes clock cells within the suprachiasmatic function in epilepsy. Curr Opin Neurol 18(2):141–145. circadian clock. Neuron 25(1):123–128. 24. Chung L (2012) Recent progress in GABAergic excitation from mature brain. Arch 6. Moore RY, Speh JC, Leak RK (2002) Suprachiasmatic nucleus organization. Cell Tissue Pharm Res 35(12):2035–2044. Res 309(1):89–98. 25. Fritschy JM (2008) Epilepsy, E/I Balance and GABA(A) Receptor Plasticity. Front Mol 7. Baroncelli L, et al. (2011) Brain plasticity and disease: A matter of inhibition. Neural Neurosci 1:5. Plast 2011:286073. 26. Lewy AJ, Lefler BJ, Emens JS, Bauer VK (2006) The circadian basis of winter depression. 8. Trevelyan AJ, Schevon CA (2013) How inhibition influences seizure propagation. Proc Natl Acad Sci USA 103(19):7414–7419. Neuropharmacology 69:45–54. 27. Leach G, Ramanathan C, Langel J, Yan L (2013) Responses of brain and behavior to 9. Choi HJ, et al. (2008) Excitatory actions of GABA in the suprachiasmatic nucleus. changing day-length in the diurnal grass rat (Arvicanthis niloticus). Neuroscience 234: J Neurosci 28(21):5450–5459. 31–39. 10. De Jeu M, Pennartz C (2002) Circadian modulation of GABA function in the rat su- 28. Falchi F, Cinzano P, Elvidge CD, Keith DM, Haim A (2011) Limiting the impact of light prachiasmatic nucleus: Excitatory effects during the night phase. J Neurophysiol 87(2): pollution on human health, environment and stellar visibility. J Environ Manage 834–844. 92(10):2714–2722. + 11. Irwin RP, Allen CN (2009) GABAergic signaling induces divergent neuronal Ca2 re- 29. Dijk DJ, et al. (2012) Amplitude reduction and phase shifts of melatonin, cortisol and sponses in the suprachiasmatic nucleus network. Eur J Neurosci 30(8):1462–1475. other circadian rhythms after a gradual advance of sleep and light exposure in hu- 12. Wagner S, Castel M, Gainer H, Yarom Y (1997) GABA in the mammalian supra- mans. PLoS ONE 7(2):e30037. chiasmatic nucleus and its role in diurnal rhythmicity. Nature 387(6633):598–603. 30. Daan S, Merrow M, Roenneberg T (2002) External time—internal time. J Biol Rhythms 13. Wagner S, Sagiv N, Yarom Y (2001) GABA-induced current and circadian regulation of 17(2):107–109. chloride in neurones of the rat suprachiasmatic nucleus. J Physiol 537(Pt 3):853–869. 31. Farajnia S, et al. (2012) Evidence for neuronal desynchrony in the aged supra- 14. Itri J, Michel S, Waschek JA, Colwell CS (2004) Circadian rhythm in inhibitory synaptic chiasmatic nucleus clock. J Neurosci 32(17):5891–5899. transmission in the mouse suprachiasmatic nucleus. J Neurophysiol 92(1):311–319. 32. Michel S, Itri J, Colwell CS (2002) Excitatory mechanisms in the suprachiasmatic nu- 15. Irwin RP, Allen CN (2007) Calcium response to retinohypothalamic tract synaptic cleus: The role of AMPA/KA glutamate receptors. J Neurophysiol 88(2):817–828. transmission in suprachiasmatic nucleus neurons. J Neurosci 27(43):11748–11757. 33. Billups D, Attwell D (2002) Control of intracellular chloride concentration and GABA 16. Bartos M, et al. (2002) Fast synaptic inhibition promotes synchronized gamma oscillations response polarity in rat retinal ON bipolar cells. J Physiol 545(Pt 1):183–198. in hippocampal interneuron networks. Proc Natl Acad Sci USA 99(20):13222–13227. 34. Yu J, Proddutur A, Elgammal FS, Ito T, Santhakumar V (2013) Status epilepticus en- 17. Van Vreeswijk C, Abbott LF, Ermentrout GB (1994) When inhibition not excitation hances tonic GABA currents and depolarizes GABA reversal potential in dentate fast- synchronizes neural firing. J Comput Neurosci 1(4):313–321. spiking basket cells. J Neurophysiol 109(7):1746–1763. 18. Freeman GM, Jr., Krock RM, Aton SJ, Thaben P, Herzog ED (2013) GABA networks 35. Michel S, et al. (2013) Mechanism of bilateral communication in the suprachiasmatic destabilize genetic oscillations in the circadian pacemaker. Neuron 78(5):799–806. nucleus. Eur J Neurosci 37(6):964–971.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1319820111 Farajnia et al. Downloaded by guest on October 2, 2021